Session library

Reimagining Creative Agencies with AI The Bad Marketing Story

Yotam Dor & Ian Buck · AI Marketers Guild · November 12, 2026

Bad Marketing co-founders Ian Buck and Yotam Dor explain how their two-person agency leverages an AI operating system to scale to $2.5 million in revenue.

ai agenciesai agents

Bad Marketing co-founders Ian Buck and Yotam Dor explain how their two-person agency leverages an AI operating system to scale to $2.5 million in revenue. In this session hosted by David Berkowitz for the AI Marketers Guild, the founders detail how they broke away from traditional holding company models to build an agency of record powered by automated business operations and personalized knowledge graphs.

Rather than relying on generative AI to produce cheap marketing collateral, Bad Marketing uses artificial intelligence to overhaul backend agency operations, client communications, and strategic planning. They explain how automating project management, invoicing, and talent coordination allows a lean senior team to outmaneuver traditional agencies burdened by billable hours and administrative overhead.

The discussion covers their proprietary AI chief of staff named "Goody," their approach to encoding creative taste into custom AI models, and why authentic human production remains their primary differentiator. Buck, Dor, and their newly appointed CTO, Matt, also share how they plan to productize and license this AI operating system to other independent creative shops.

Key moments

[05:15] Why Did Bad Marketing Leave Holding Companies to Build an AI-Native Creative Agency?

Ian Buck explains that Bad Marketing was founded because traditional advertising holding companies like Publicis and IPG lacked the organizational appetite and structural ability to innovate rapidly with artificial intelligence. Having previously sold an agency to Publicis and worked inside IPG, Buck and co-founder Yotam Dor realized that legacy agencies were too encumbered by established hierarchies and billable-hour structures to fundamentally rethink their business workflows around emerging technology.

When the founders launched Bad Marketing, most of the industry conversation was fixated on generative AI and imperfect visual outputs like six-fingered hands. Buck and Dor recognized that generative image and text quality would inevitably improve, so they focused instead on the deeper operational opportunity: reimagining agency processes, workflows, and business models from the ground up to operate with unprecedented efficiency.

Notable Quotes:

"While we were within these giant companies, what we recognized was there is no real appetite for innovation in companies that huge. It's just really hard to do." — Ian Buck [05:37]

"Really, the for us the generative side wasn't the most interesting side. It was like what about like the business side, processes and and you know, workflows, what is the opportunity there?" — Ian Buck [07:13]

[10:41] How Can a Two-Person Agency Generate $2.5 Million in Revenue Using AI?

Ian Buck reports that Bad Marketing scaled its annual revenue to nearly $2.5 million while serving as the agency of record for major enterprises like Jamieson Vitamins and RBC with only two full-time employees. In a conventional agency model, servicing accounts of that scale and revenue would require 10 to 12 full-time staff members, making talent the single largest operational expense.

By embedding custom AI tools and automation into the back office, the agency inverted the traditional cost structure. Because AI handles routine project management, contracting, and coordination, the founders operate with virtually zero payroll overhead beyond themselves. This lean structure gives the agency an entirely different economic model and a sustainable competitive advantage over legacy firms.

Notable Quotes:

"Within 3 years we got our annual revenue to about just under 2.5 million. And what was really interesting for us there was in our past two agencies and in basically every other agency in the world, you would need about 10 to 12 full-time employees to run a an agency at that revenue rate" — Ian Buck [11:33]

"Your real cost is talent. And so when you can run an agency with two people versus 10 to 12, that's not even a competitive advantage, that is an entirely different thing, right?" — Ian Buck [12:28]

[13:36] Why Does Having Senior Founders in Every Meeting Save Time for Both Agency and Client?

Ian Buck explains that staffing every client touchpoint exclusively with senior founders dramatically compresses project timelines by enabling instant decision-making in the room. When clients surveyed Bad Marketing after three years of work, their primary praise was that the agency's most senior decision-makers were present in every single meeting.

This structure eliminates the traditional agency layers where junior account representatives take notes, carry feedback back to internal leadership, draft proposals, and wait days for sign-off. Crucially, Buck notes this saves massive amounts of time on the client side as well, because client project managers and junior staff can make immediate progress without scheduling extra internal briefings or status checks for their own executives.

Notable Quotes:

"What it turns out is we save not just us a ton of time but it saves them a massive ton of time to have the most senior people in every meeting." — Ian Buck [14:24]

"We basically as the most senior people can solve problems in that room, right? There's no like taking stuff back to our team to discuss and come back with a plan and go, you know, like this." — Ian Buck [14:41]

[20:55] How Does Bad Marketing Execute Large Production Campaigns With Minimal Core Staff?

Yotam Dor explains that Bad Marketing executes high-volume client deliverables by combining in-house strategic direction with a trusted, distributed network of specialist creators. The agency's core team provides the overarching ideas, client management, and brand strategy, while vetted external directors, writers, designers, and production crews handle execution at scale.

To orchestrate this distributed network without hiring a layer of project managers, the founders automated their operational pipeline directly using command-line AI tools and custom software. Dor emphasizes that once operational workflows are codified in software, agency capacity expands exponentially because repeatable coordination tasks never need to be performed manually again.

Notable Quotes:

"Really what we use is a distributed network of creators that allows us to facilitate a lot of the client work so that the ideas and the strategy and just the experience is what the clients get direct impact from." — Yotam Dor [21:26]

"Every single day, every single hour, we were adding new ways of doing things that we would never need to repeat ourselves again. And once you and what what's amazing about code is as you all know once you've done it once then it starts to allow you to go exponential" — Yotam Dor [22:16]

[23:42] Why Does AI Accelerate the Shift From Billable Hours to Project-Based Pricing?

Ian Buck argues that the billable hour model is structurally broken in an AI-powered world because artificial intelligence rapidly compresses production time, penalizing efficient agencies under hourly billing. Having moved away from billable hours in the late 2000s while managing social media accounts, Buck and Dor structure all client engagements on project-based pricing to align compensation directly with value and strategic outcomes.

Yotam Dor adds that offloading administrative back-office work onto AI machines frees the agency to focus entirely on human judgment, strategic insight, and creative taste. By pricing projects rather than hours, the agency captures the full financial benefit of its efficiency gains while providing clients with senior-level attention and superior outcomes.

Notable Quotes:

"So we very early on started thinking about that from a how do you move from billable hours which doesn't work in a compressed time to project." — Ian Buck [25:07]

"From a business perspective, it's actually a really great way to unlock a new way of charging for value that billable hours does not give you especially now that things are accelerated by AI which is great at compressing time." — Ian Buck [26:14]

"We're so focused on taking all of the back office work and putting it onto the machines so that our focus, our insights, our knowledge and and our taste could be could be, you know, directly impacting our clients." — Yotam Dor [28:00]

[29:13] How Can an Agency Encode Creative Taste and Judgment Into a Custom AI Knowledge Graph?

Ian Buck explains that Bad Marketing encodes its creative taste and strategic frameworks into a private knowledge graph rather than relying on generic, off-the-shelf AI models. When the agency was founded, they worked with an engineer to build a custom graph populated with successful pitch decks, strategic brand frameworks, voice memos, and recorded founder conversations that encapsulated their creative philosophy.

Connecting the agency's data directly to large language models allows the system to act as a personalized creative force multiplier. When deploying this operating system to other agencies, Buck emphasizes that the system requires a custom knowledge graph for each shop, enabling creative leaders to interact with an AI that mirrors their unique judgment and aesthetic values.

Notable Quotes:

"On top of that, you're also dipping into essentially that knowledge graph that is specifically you. So that's the idea of when we take our operating system and put it on a new agency. It's a custom build every time" — Ian Buck [31:33]

"Really the important part is exactly what you're talking about is how do you get your brain, your taste, your judgment into that system so it's for a reflection of you." — Ian Buck [32:18]

"In that moment where you ask a question and suddenly you get exactly the right answer right in in a matter of seconds that would have taken you, you know, minutes or maybe hours to to figure out yourself. That is that aha moment" — Yotam Dor [35:24]

[36:20] What Agency Administrative Tasks Can an AI Chief of Staff Like Goody Automate?

Ian Buck details how their custom AI chief of staff, dubbed "Goody," automates the time-consuming administrative tasks that typically bog down agency operators. Interfacing through both voice and chat, Goody manages vendor invoicing, tracks cash flow, matches contractor bills against original budget estimates, schedules meetings via automated email back-and-forth, and delivers a daily morning briefing to the founders.

Goody also handles the full lifecycle of contractor management by drafting freelancer agreements, dispatching them for digital signature, chasing pending sign-offs, and archiving executed contracts. Crucially, the system includes strict human-in-the-loop safeguards: Goody cannot issue payments, alter contract terms, or finalize agreements without explicit approval from the founders.

Notable Quotes:

"Goodie is our chief of staff, you know that we started vibe coding back in the day and just figuring out what can we make that kind of helps our process work a lot better." — Ian Buck [36:58]

"It's doing all of our freelancer agreements. So it's actually drafting the agreements. It's sending it out for signature. It's chasing those signatures and then it's filing the signed documents." — Ian Buck [42:29]

"Never pays anyone without us given the okay. Agreements that need signature need to be kind of checked by us. So there, you know, you retain those moments to retain the control." — Ian Buck [46:21]

[39:30] How Does Administrative AI Automation Reduce Cognitive Load for Creative Teams?

Ian Buck explains that the primary value of AI workflow automation is eliminating cognitive interruptions and mental context-switching rather than merely saving raw hours. During a production test involving a day-long film shoot with 28 separate freelance crew members, automating the contracting, invoicing, and tracking processes through Goody saved approximately 40 hours of administrative work per month.

Buck emphasizes that the psychological relief far exceeded the 40 saved hours. When agency operators must constantly open emails to resolve billing discrepancies or chase missing signatures, their creative flow is fractured, forcing them to catch up on high-value creative work late at night. Automating administrative busywork preserves mental energy and allows operators to sustain deep strategic focus.

Notable Quotes:

"We actually measured that on a test with 28 people and we got back that it saves about a week every month. So 40 hours right in savings by automating that process with goodie." — Ian Buck [40:13]

"It's actually those shifts, you know, the cognitive interruptions that are the real savers there." — Ian Buck [41:09]

"You realize like it's not just a time savings that you're getting. You're getting this like very real cognitive savings that allows you to really hyperfocus or or really just do the work instead of all the other stuff around it." — Ian Buck [41:39]

[47:32] How Does Bad Marketing Use AI to Accelerate Client Briefs and Creative Proposals?

Ian Buck explains that Bad Marketing bypasses lengthy, weeks-long corporate briefing processes by conducting focused 30-minute discovery calls and feeding the requirements directly into their AI knowledge tool. The AI proposes initial angles and creative structures that Buck directs and iterates against, functioning like a fleet of junior strategists to produce a complete creative proposal in a single day.

Despite being able to turn around comprehensive proposals in 24 hours, Buck notes that the agency deliberately waits a week before presenting work to clients. Because clients are accustomed to proposals requiring weeks of labor, delivering high-level creative work overnight creates cognitive dissonance, leading clients to assume that rapidly generated work must be low quality or lacking in effort.

Notable Quotes:

"We've actually asked them to not do a brief and could we just have like a halfhour chat instead. And so, in that chat, we just ask a few questions." — Ian Buck [48:15]

"What we found is we can actually build that in a day and so we've tried this I've pretended like they wanted us to present the next Okay, so we've built it all out within a day, but then we actually wait." — Ian Buck [51:21]

"They wouldn't value it because they wouldn't actually believe that it's possible. So they would say this is not good work because there's no way it good work takes more time than this, right?" — Ian Buck [52:21]

[52:47] Why Does Bad Marketing Avoid Using Generative AI for Final Client Deliverables?

Ian Buck states that Bad Marketing deliberately avoids using generative AI in the production of final client assets, choosing instead to execute visual and video work with human creative professionals. As generative AI makes digital content creation ubiquitous and lowers production barriers across the industry, Buck argues that human authenticity and craftsmanship will become the defining differentiator for premium brands.

While the agency heavily integrates AI across strategy, project coordination, research, and internal workflows, the final consumer-facing deliverables are produced by real directors, cinematographers, and artists. Maintaining this clear separation between operational AI efficiency and human-crafted production ensures high creative standards while avoiding intellectual property and copyright complications for client brands.

Notable Quotes:

"We believe that actually there's a huge advantage for brands to be not using generative AI in the production process. So that's not really part of what we do. We have a unique production methodology that allows us to create everything with real people." — Ian Buck [53:13]

"We believe fundamentally that that kind of authenticity will be one of the differentiators for brands moving forward. As the generative stuff raises the entire ocean, right? And if everyone can make that then what does how do how does a brand differentiate itself" — Ian Buck [53:31]

[54:48] How Should AI-Powered Agencies Manage Client Data Security, Privacy, and Workflow Pace?

Yotam Dor advises that the fundamental first step for any agency using AI is turning off model training on all enterprise tools to prevent sensitive client data from leaking into public foundational models. Agencies must remain adaptable, customizing their data pipelines and storage practices to match the specific Master Services Agreement (MSA) and compliance requirements of each corporate client.

Ian Buck adds that agencies adopting AI must handle client workflow differences with patience and care. While an AI-native agency can operate at extreme speed, enterprise clients remain constrained by internal legal checks, brand approvals, and corporate protocols. Rather than forcing clients into uncomfortable speeds, the agency should identify targeted internal efficiencies that save client teams time while respecting their institutional pace.

Notable Quotes:

"What you really realize very quickly is that they can only be so much quicker. They are fundamentally kind of like handcuffed by the processes that they need to have in place for themselves, right? And we can't really change that. So we essentially just adapt to them" — Ian Buck [55:12]

"The first thing I would suggest is anyone working with AI today ensure that you've switched off training on any of your models would be step one and making sure that you're not a civ for all of your client data" — Yotam Dor [56:28]

"What we do is we just work in our own processes to map back against whatever the MSA is with with that client, which makes it, you know, fun for Matt as he builds this with us. But it's been one of the greatest learnings for us is that you need to be a reflection of your client." — Yotam Dor [57:22]